Q3C, Quad Tree Cube -- The new Sky-indexing Concept for Huge Astronomical Catalogues and its Realization for Main Astronomical Queries (Cone Search and Xmatch) in Open Source Database PostgreSQL

Abstract

, Oleg BartunovSternberg Astronomical Institute, Universitetskiy pr. 13, 111992,Moscow, RussiaAbstract. In this paper we present Q3C (Quad Tree Cube), a new skyindexing scheme and its implementation for the open source databasePostgreSQL.Wehaveimplementedoursky-partitioningschemeasaload-able module for PostgreSQL and developed a very simple SQL interfaceformainastronomicalqueries: conesearch, variousspatialsearchesonthesphereandcross-matchesofcatalogues. Theperformanceisveryhighandallows to work easily with the largest existing catalogues (USNO-A/B,2MASS, SDSS). We propose PostgreSQL and our sky-indexing scheme asopen source solution for any VO-services dealing with huge catalogues.1. IntroductionIn last few years, several projects have provided to community with a set ofvery large catalogues consisting of up to billion of objects. Simple access meth-ods are no longer suitable to work with that amount of data and databasesmust be used instead. But even the most advanced databases are lacking formethods to effectively work with spherical coordinates and astronomical querieslike cone searches and cross-matches. So, special astronomical indexes and sky-partitioningschemes were invented for this purpose. The first successful attemptwas HTM – hierarchical triangular mesh (Kunszt 2000). This sky partitioningscheme has been used in MS SQL to provide the interfaces for the astrono-mical queries in the SDSS project. However, HTM has several disadvantages(O’Mullane 2000): first, HTM is a too complicated scheme which significantlylimits its performance (in particular, in those cases with a high depth of seg-mentation); second, HTM was specifically developed for Microsoft SQL server,which is not open source. That was the motivation for the development of anew simple, fast and powerful sky-indexing scheme for open source databasePostgreSQL.